FLEXIBLE NITROGEN DIOXIDE GAS SENSOR BASED ON TUNGSTEN TRIOXIDE NANOPARTICLES COATED CARBON NANOTUBES-GRAPHENE OXIDE HYBRID AND METHOD FOR MANUFACTURING THE SAME

Abstract

A flexible nitrogen dioxide sensor based on tungsten trioxide nanoparticles-loaded multi-walled carbon nanotubes-reduced graphene oxide (WO.sub.3 NPs-loaded MWCNTs-RGO) hybrid on a polyimide/polyethylene terephthalate substrate. A viscous gel of the hybrid materials can be prepared by the assistance of -terpineol. The fabricated sensor shows excellent sensing performance toward NO.sub.2 which may have a maximum response of 17% (to 5 ppm), a limit of detection of 1 ppm, and relatively short response/recovery time (7/15 min). The sensor may exhibit excellent mechanical flexibility and sensing properties at room temperature without any significant performance degradation even at a curvature angle of 90 and after 10.sup.6 times of bending/relaxing processes. Low cost, light weight and mechanical robustness of the proposed WO.sub.3 NPs-MWCNTs-RGO hybrid based sensor can be a promising element for the development of flexible NO.sub.2 gas sensors having higher performance.

Claims

1. A flexible nitrogen dioxide (NO.sub.2) gas sensor based on a tungsten trioxide nanoparticles (WO.sub.3 NPs) coated carbon nanotubes (CNTs)-graphene oxide (GO) hybrid, the flexible gas sensor detecting NO.sub.2 gas being manufactured by using a tungsten trioxide nanoparticles (WO.sub.3 NPs)-loaded multi-walled carbon nanotubes (MWCNTs)-reduced graphene oxide (RGO) hybrid where multi-walled carbon nanotubes (MWCNTs), reduced graphene oxide (RGO), and tungsten trioxide nanoparticles (WO.sub.3 NPs) are mixed at a proper mixing ratio.

2. The flexible nitrogen dioxide (NO.sub.2) gas sensor of claim 1, wherein the multi-walled carbon nanotubes (MWCNTs) and the reduced graphene oxide (RGO) are mixed with an assistance of -terpineol to prepare a hybrid mixture solution, before adding the tungsten trioxide nanoparticles (WO.sub.3 NPs) thereto.

3. The flexible nitrogen dioxide (NO.sub.2) gas sensor of claim 1, wherein the multi-walled carbon nanotubes (MWCNTs), the reduced graphene oxide (RGO), and the tungsten trioxide nanoparticles (WO.sub.3 NPs) in the hybrid are mixed at a ratio of 3:1:2 by weight.

4. The flexible nitrogen dioxide (NO.sub.2) gas sensor of claim 2, wherein the hybrid mixture solution is dropped on a space between two gold (Au) electrodes, away from each other at a predetermined distance, deposited on a polyimide (PI)/Si substrate and dried to prepare the sensor.

5. The flexible nitrogen dioxide (NO.sub.2) gas sensor of claim 4, wherein the gas sensor with the mixing ratio shows a maximum response value of 17% at an annealing temperature of 200 C.

6. The flexible nitrogen dioxide (NO.sub.2) gas sensor of claim 4, wherein the gas sensor exhibits a limit of detection (LOD) of 1 ppm and a detection range of 1 to 25 ppm.

7. The flexible nitrogen dioxide (NO.sub.2) gas sensor of claim 4, wherein the gas sensor exhibits sensing performance after certain times of bending/relaxing cycles and at a certain curvature angle.

8. A method for manufacturing a flexible nitrogen dioxide (NO.sub.2) gas sensor based on a tungsten trioxide nanoparticles (WO.sub.3 NPs) coated carbon nanotubes (CNTs)-graphene oxide (GO) hybrid, the method comprising: preparing a starting solution by mixing multi-walled carbon nanotubes (MWCNTs) and synthesized graphene oxide (GO) powders with an assistance of -terpineol; adding tungsten trioxide nanoparticles (WO.sub.3 NPs) to the starting solution at a predetermined mixing ratio with the multi-walled carbon nanotubes (MWCNTs) and the graphene oxide (GO); dropping the starting solution on a place between two gold (Au) electrodes, away from each other at a predetermined distance, deposited on a polyimide (PI)/Si substrate made of a PI tape and a Si substrate, and drying, and then performing annealing; and removing the polyimide (PI) tape from the PI/Si substrate and transferring the PI tape on to a polyethylene (PET) substrate.

9. The method of claim 8, wherein the multi-walled carbon nanotubes (MWCNTs), the graphene oxide (GO), and the tungsten trioxide nanoparticles (WO.sub.3 NPs) are mixed at a ratio of 3:1:2 by weight.

10. The method of claim 8, wherein the polyimide (PI)/Si substrate on which the gold (Au) electrodes are deposited and the starting solution is dried is annealed at 200 C.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0025] FIG. 1(a) is a schematic diagram illustrating a fabricated gas sensor according to an embodiment of the present disclosure.

[0026] FIG. 1(b) is an optical image of a fabricated gas sensor according to an embodiment of the present disclosure.

[0027] FIG. 2(a) shows an SEM image of a WO.sub.3 NPs-loaded MWCNTs-RGO hybrid where WO.sub.3 NPs, MWCNTs, and RGO are mixed at a mixing ratio of 3:0.5:1 (S1).

[0028] FIG. 2(b) shows an SEM image of a WO.sub.3 NPs-loaded MWCNTs-RGO hybrid where WO.sub.3 NPs, MWCNTs, and RGO are mixed at a mixing ratio of 3:1:2 (S2).

[0029] FIG. 2(c) shows an SEM image of a WO.sub.3 NPs-loaded MWCNTs-RGO hybrid where WO.sub.3 NPs, MWCNTs, and RGO are mixed at a mixing ratio of 3:2:3 (S3).

[0030] FIG. 2(d) shows an SEM image of a WO.sub.3 NPs-loaded MWCNTs-RGO hybrid S2 after several times of bending/relaxing.

[0031] FIG. 3(a) shows a TEM image of S2 hybrid at low magnification.

[0032] FIG. 3 (b) shows a TEM image of S2 hybrid at high magnification.

[0033] FIG. 3(c) shows an HRTEM image at an interface between WO.sub.3 NP and MWCNT.

[0034] FIG. 3(d) shows an EDS spectrum of S2 hybrid.

[0035] FIG. 4 shows XRD patterns of a WO.sub.3 NPs-loaded MWCNTs-RGO hybrid at different annealing temperatures.

[0036] FIG. 5(a) shows Raman spectra of pure MWCNTs, RGO, and a MWCNTs-RGO hybrid.

[0037] FIG. 5(b) shows Raman spectra of S1, S2, and S3 hybrids.

[0038] FIG. 6 shows BET results of pure WO.sub.3, a WO.sub.3-MWCNTs hybrid, and a WO.sub.3-MWCNTs-RGO hybrid.

[0039] FIG. 7 shows response values of S1, S2, and S3 sensors at different annealing temperatures toward 5 ppm NO.sub.2 gas.

[0040] FIG. 8 shows real time resistance changes of S1, S2, and S3 sensors toward NO.sub.2.

[0041] FIG. 9 shows comparison of response values among pure MWCNTs, a WO.sub.3-MWCNTs hybrid, and a WO.sub.3 NPs-loaded MWCNTs-RGO hybrid.

[0042] FIG. 10 shows transient response values of S2 hybrid toward 5 ppm NO.sub.2 gas at different bending angles.

[0043] FIG. 11 shows response/recovery time characteristics of S2 hybrid toward 5 ppm NO.sub.2 gas.

[0044] FIG. 12(a) shows sensor response value vs. NO.sub.2concentration curve at different bending angles.

[0045] FIG. 12(b) shows sensor response value vs. NO.sub.2concentration curve after several times of bending/relaxing.

[0046] FIG. 13 shows humidity effect on a sensor for NO.sub.2concentration of 5 ppm.

[0047] FIG. 14 is a selectivity graph of a fabricated sensor toward 5 ppm NO.sub.2 and 1000 ppm test gases.

DETAILED DESCRIPTION

[0048] Hereinafter, preferred example embodiments of a flexible NO.sub.2 gas sensor based on a WO.sub.3 NPs coated CNTs-GO hybrid and a method for manufacturing the same according to the present disclosure are described in detail with reference to the accompanying drawings. However, it should be noted that the present disclosure is not limited to the embodiments described below but may come with a diversity of embodiments. The embodiments are provided only to complete the present disclosure and help those skilled in the art fully understand the scope of the present disclosure.

[0049] A flexible NO.sub.2 gas sensor based on a WO.sub.3 NPs coated CNTs-GO hybrid and a method for manufacturing the same according to the present disclosure are described in detail. The fabrication of a flexible NO.sub.2 gas sensor according to the present disclosure is described through the following experiment.

1. Experiment

1.1 Materials Synthesis and Sensor Fabrication

[0050] All of the chemicals used in the synthesis process were of analytical grade purchased from Sigma Aldrich, Dongwoo Fine-Chem., and Dae Jung Chem. & Inds. Co. Ltd., and were used without further purification.

[0051] Synthesis of WO.sub.3NPs and graphene oxide (GO): Tungstatedihydrate (Na.sub.2WO.sub.4.2H.sub.2O) and cetyltrimethylammonium bromide (CTAB) were used as precursor and surfactant, respectively. In a typical process, 10 mL of CTAB (0.5 M) and 10 mL of Na.sub.2WO.sub.4.2H.sub.2O (1.5 M) aqueous solution were added to 20 mL of de-ionized (DI) water using vigorous stirring. 1 mL of HCl (3 M) was subsequently added drop-wise to the solution to obtain a pH level around 3. The as-prepared solution was then transferred into a sealed Teflon autoclave and heated at 120 C. for 12 h. The final product was collected after several times of washing with ethanol-DI water (1:1) and dried overnight at 60 C. GO solution was synthesized by modified Hummer's method as describes in elsewhere(W. S. Hummers Jr. et al., Preparation of graphitic oxide, J. Am. Chem. Soc. 80 (1958), 1339-1339; D. T. Phan et al., Photodiodes based on graphene oxide-silicon junctions, Solar Energy 86 (2012) 2961-2966). GO solution was dried at 45 C. for 48 h to obtain fine GO powders.

[0052] Synthesis of a WO.sub.3 NPs-loaded MWCNTs-RGO hybrid: Commercial MWCNTs (=4.5 to 5 nm; 1=3 to 6 m) and synthesized GO powders were mixed with an assistance of -terpineol using sonication treatment for 1 h, to prepare a starting solution. Thereafter, 3 mg of WO.sub.3 NPs powders were added to the starting solution and subjected to sonication treatment for another 1 h. To obtain an optimum material ratio, three different hybrid samples were prepared by varying the amounts of MWCNTs and GO to WO.sub.3 (WO.sub.3:MWCNTs:GO=(a) 3:0.5:1; (b) 3:1:2; (c) 3:2:3).

[0053] Sensor fabrication: To fabricate a sensor, a commercial PI tape was attached on a Si substrate. Two finger electrodes of gold (distance: 100 m) were deposited on the top of a PI/Si substrate using photolithography and radio frequency magnetron sputtering. The as-prepared hybrid was then drop casted between the finger electrodes and placed on a hot plate at 100 C. for drying. Afterward, each sample was annealed at different temperatures (100, 150, 200, and 250 C.) for 1 h. For better clarity, sensors were labeled as S1 (WO.sub.3 NPs-MWCNTs-RGO =3:0.5:1); S2 (WO.sub.3 NPs-MWCNTs-RGO=3:1:2) and S3 (WO.sub.3 NPs -MWCNTs-RGO=3:2:3). Finally, the PI tape was carefully peeled-out from the Si substrate and transferred on to a PET substrate. The schematic diagram and the optical image of the fabricated sensor are shown in FIG. 1(a) and FIG. 1(b), respectively.

1.2 Characterization

[0054] Phase transition analysis was carried out by an X-ray diffractometer (XRD) (Ultima IV, Rigaku Corporation) with Cu K (=0.154056 nm) radiation and a 2 scanning range of 10 to 70. The surface morphology and elemental characterizations of the as-prepared hybrids were examined by field emission scanning electron microscopy (FESEM, JEOL-JSM-7600F), transmission electron microscopy (TEM, JEOL JEM-2100F), high-resolution TEM (HRTEM) and energy dispersive spectroscopy (EDS, JEOL JEM-2100F). Raman spectra were acquired through a WITec spectrometer with 532 nm laser excitation in order to detect possible structural properties and quality of the synthesized hybrid materials. The Brunauer Emmett and Teller (BET) analysis of pure MWCNTs, WO.sub.3-MWCNTs, and WO.sub.3-MWCNTs-RGO was measured by nitrogen adsorption at 77 K temperature to observe specific surface area (SSA(.sub.BET)) of each material.

[0055] The gas sensing characterizations were carried out at room temperature in an open air environment. A computerized mass flow controller system (GMC 1200 ATOVAC Co., Ltd.) was used to vary the NO.sub.2 gas concentration. A gas mixture (synthetic air and NO.sub.2) was delivered on the top of a sensor device at a constant flow rate of 50 standard cubic centimeters per minute (sccm) with different NO.sub.2 concentrations. Gas concentration was controlled and measured by the following equation 1:

[00001]${\mathrm{Gas}}_{\mathrm{com}}\left(\mathrm{ppm}\right)=\frac{{\mathrm{Flowrate}}_{\mathrm{air}}+{\mathrm{Flowrate}}_{\mathrm{gas}}}{\mathrm{Totalflowrate}}$

A sensor response value was calculated by the following equation 2:

[00002]$S\left(\%\right)=\frac{R_a-R_g}{R_a}100,$

[0056] where S (%) denotes a sensor response value in percentage, R.sub.a is the resistance of the sensor in air, and R.sub.g is the resistance after exposure to a certain amount of NO.sub.2. The response time and recovery time of the sensor were defined as the time taken to reach 90% of the total resistance change.

2. Results and Discussions

2.1 Crystal Structure and Morphology

[0057] FIG. 2 shows FESEM images of WO.sub.3 NPs-loaded MWCNTs-RGO hybrids (samples S1, S2 and S3 annealed at 200 C.). All of the observed samples include irregularly shaped WO.sub.3 nanoparticles, long isolated MWCNTs, and thin RGO nanosheets. A small amount of aggregation was frequently observed in the samples that can probably be accounted for the presence of different materials with high density. WO.sub.3 nanoparticles were more visible in sample S1 (FIG. 2(a)) compared to other samples (S2 (FIGS. 2(b)) and S3 (FIG. 2(c))). However, a large amount of aggregation occurred in sample S3 due to the presence of large numbers of RGO and MWCNTs that unsurprisingly covered the WO.sub.3 nanoparticles. The samples were annealed at different temperatures (100 to 250 C.) and investigated again. The variation in the annealing temperature did not reveal any changes in the morphology of the hybrid samples. Later on, samples S1, S2 and S3 (annealed at 200 C.) were bended-relaxed repeatedly up to 10.sup.7 cycles and analyzed by FESEM. FIG. 2(d) represents the FESEM image of sample S2 after 10.sup.7 times of bending and relaxing. No significant deformation or degradation was observed after several times of bending and relaxing. This phenomenon might be attributed to exceptional mechanical robustness and outstanding flexibility properties of the MWCNTs and RGO in the hybrids and superior bending between the hybrids and PI substrate.

[0058] For detailed morphological investigations, TEM analysis was carried out at different magnifications and shown in FIG. 3. FIG. 3(a) affirms the decoration of tiny sized WO.sub.3 NPs on transparent RGO sheets and long MWCNTs. The average size of WO.sub.3 NPs was estimated to 20 nm from TEM observations. However, WO.sub.3 NPs with irregular shape was also observed, which can be attributed to the aggregation among the small sized particles. FIG. 3(b) shows the morphology of the as-prepared hybrid at high magnification. FIG. 3(c) presents the HRTEM image which indicates rightly fitted WO.sub.3 NPs on MWCNTs with continuity of lattice fringes of WO.sub.3 NPs and MWCNTs. The measured spacing between adjacent lattice fringes is 0.37 nm, corresponding to (200) plane of WO.sub.3. FIG. 3(d) represents the elementary analysis of the hybrid sample (S2) and confirms the presence of tungsten (W), carbon (C), and oxygen (O). This result presents the formation of high purity hybrid sample.

[0059] XRD was carried out to analyze the crystalline structure of S2 hybrid formed as crystals, at different annealing temperatures. FIG. 4 shows the well-structured crystalline nature of the synthesized hybrid. The different diffraction peaks appearing at 2=14.11, 22.89, 26.87, 28.30, 36.67, and 62.33 corresponds to the hexagonal WO.sub.3 (100), (001), (101), (200), (201), and (401) planes, respectively (h-WO.sub.3, ICDD: 01-075-2187). Additionally, characteristic diffraction peak at 2=18.1 corresponds to RGO (002) plane, suggesting the partial reduction of GO. Furthermore, diffraction peaks at 2=43.66, 55.35, and 76.77 appeared on carbon (101), (004), and (110) planes, respectively (C, ICDD: 00-023-0064). No significant changes or shifts in peak position were observed in the spectrum after different annealing temperatures. However, some impurities peaks were removed at higher temperature leading to the purity of hybrid materials. The intensity of the peak of RGO at 250 C. was reduced as compared to 200 C. indicating low RGO contaminants. This ultimately causes the reduction in SSA of RGO.

[0060] Raman spectroscopy was carried out to study the order/disorder of hexagonal carbon structure and the effect of WO.sub.3 thereon. FIG. 5(a) represents Raman spectra of pure MWCNTs, RGO, and a MWCNTs-RGO hybrid. The D peaks of pure MWCNTs, RGO, and MWCNTs-RGO hybrid were observed between 1350 and 1355 cm.sup.1. Additionally, the G and 2D peaks of MWCNTs, RGO, and MWCNTs-RGO hybrid were observed between 1590 and 1598 cm.sup.1 and between 2689 and 2697 cm.sup.1, respectively. The low I.sub.d/I.sub.g ratio (0.90) of MWCNTs may be attributed to low disorder in carbon network. On the other hand, in RGO, high I.sub.d/I.sub.g ratio (1.15) and wide intensity peaks were observed, which may be ascribed as higher level of disorder between RGO flakes. The MWCNTs-RGO hybrid exhibits the lowest I.sub.d/I.sub.g ratio (0.87) with wide intensity peaks. The wide intensity peaks may be explained as disorder between MWCNTs and RGO hexagonal structure. Additionally, Raman spectra of WO.sub.3-MWCNTs-RGO hybrids (S1, S2, and S3) were also investigated and shown in FIG. 5(b). The lower frequency band located at 260 cm.sup.1 may be attributed to the W-O-W bending vibrations, whereas the peaks observed at 701 and 803 cm.sup.1 may be assigned to the stretching modes of the W-O-W bonds(A. Esfandiar et al., Pd-WO3/reduced graphene oxid hierarchical nanostructures as efficient hydrogen sensors, Int. J. Hydrogen Energy 39 (2014) 8169-8179). A little increase in I.sub.d/I.sub.g ratio (0.88) was observed in samples S2 and S3, compared to a MWCNTs-RGO hybrid which might be caused by WO.sub.3. Sample S1 exhibits a higher I.sub.d/I.sub.g ratio that may be caused due to the presence of high amount of WO.sub.3.

[0061] FIG. 6 shows the BET analysis of pure WO.sub.3, WO.sub.3-MWCNTs, and WO.sub.3-MWCNTs-RGO (S2, annealed at 200 C.). It clearly reveals that WO.sub.3-MWCNTs-RGO has a larger specific surface area (SSA) of 87.23 m.sup.2/g, whereas pure WO.sub.3 and WO.sub.3-MWCNTs have specific surface areas of 8.34 m.sup.2/g and 81.71 m.sup.2/g, respectively. When MWCNTs were added to pure WO.sub.3, the addition increases the SSA (approximately 10 times), which might be attributed to high porosity of MWCNTs. Furthermore, when RGO was added to WO.sub.3-MWCNTs, the addition further increases the SSA, which might be caused by the presence of the basal plane of RGO.

2.2 Gas Sensing Properties

[0062] The gas sensing properties of fabricated sensors were carried out in an open air environment at room temperature (20 C.). NO.sub.2 is an oxidizing gas, which captures electrons and subsequently increases or decreases conductance of the sensing layer. In this particular case, when NO.sub.2 gas interacts with the sensing layer, the interaction increases the conductance suggesting p-type behavior of the sensing layer (G. Lu et al., Reduced graphene oxide for room-temperature gas sensors, Nanotechnology 20 (2009) 445502-445511). The probable sensing mechanism of the fabricated sensor is explained in FIG. 1(a). When the sensor was placed in an open air environment, oxygen molecules enter and capture electrons from the surface of the sensing layer and leave oxygen absorbents (O.sub.2.sup.). When NO.sub.2 molecules enter and interact with the sensing layer surface, they take electrons, are dissociated in the foam of NO, and leave oxygen absorbents (O.sub.2.sup.). The O.sub.2.sup.anion then becomes an active site to adsorb NO.sub.2molecules. Subsequently, NO may not convert to NO.sub.2 again after reacting with half of the O.sub.2 molecules. This phenomenon continuously happened after NO.sub.2 exposure.

[0063] FIG. 7 shows the response value variations of samples S1, S2, and S3 checked in terms of different annealing temperatures toward 5 ppm NO.sub.2 concentration. Sample S2 showed a maximum response value of 17% at an optimum annealing temperature of 200 C. This may be attributed to the high specific surface area of the synthesized material and formation of the depletion layer through p-n junctions between p-type MWCNTs/RGO and n-type WO.sub.3 NPs. This depletion layer with excess of charges may play an important role to increase the sensor response value (S. Srivastava et al., Faster response of NO2 sensing in graphene-WO3 nanocomposites, Nanotechnology 23 (2012) 205501-205507; J. S. Lee et al., WO.sub.3 nanonodule-decorated hybrid carbon nanofibers for NO2 gas sensor application, J. Mater. Chem. A 1 (2013) 9099-9106). After the exposure of NO.sub.2, a number of charges were transferred from the specific region for NO.sub.2molecules resulting in a dramatic increase in the sensor response. Annealing temperature also plays a significant role to enhance the sensor performance. At lower temperatures (100 and 150 C.), probably, GO was not completely reduced. In contrast, at a higher temperature (250 C.), RGO might be decomposed slightly (XRD shows a minimum decrease at intensity peak) and might cause reduction in SSA. It was supposed that at an optimum annealing temperature of 200 C., -terpineol was fully removed and RGO exhibits maximum SSA.

[0064] FIG. 8 shows the real time resistance changes of all the sensors (S1, S2, and S3, annealed at 200 C.) in terms of NO.sub.2 gas concentration. It was observed that samples S1 and S3 have the highest and the lowest resistance values, respectively. This phenomenon might be caused by the variation in amounts of MWCNTs and RGO. The presence of higher amounts of MWCNTs and RGO in S3 resulted in higher conductivity, opposed to forming balanced p-n junctions, and lower selectivity to NO.sub.2 gas. On the contrary, sample S1 contained a higher amount of WO.sub.3 compared to the other samples that may create (O.sub.ads.sup. or O.sup.2.sub.ads adsorbents on the sensing surface, and prevented NO.sub.2 molecules from reacting at lower temperatures. In comparison to samples S1 and S3, S2 was endowed with suitable p-n junctions that created adequate active sites to adsorb a maximum number of gas molecules. In addition, larger SSA played an important role in response enhancement.

[0065] FIG. 9 shows the response properties of pure MWCNTs and WO.sub.3-MWCNTs hybrid, compared to WO.sub.3-loaded MWCNTs-RGO hybrid sensor (sample S2), in order to support the BET results. It was clearly observed that S2 sensor exhibits a highest response value along with a shorter recovery time, compared to pure MWCNTs and WO.sub.3-MWCNTs hybrid sensor. The high SSA and enhanced charge transfer pathway were provided by the WO.sub.3 NPs, RGO sheets, and MWCNTs network which facilitated the adsorption-desorption kinetics during sensor characterizations, hence showed better sensor performance. Moreover, a low response value and partial recovery of the pure MWCNTs sample might be attributed to the absence of p-n junction and high bending energy between carbon and NO.sub.2 molecules(J. Li et al., Carbon nanotube sensors for gas and organic vapor detection, Nano Lett. 3 (2003) 929-933).

[0066] To investigate the reliability and mechanical flexibility, the fabricated sensor sample S2 was evaluated at different curvature angles (0 to 90 at room temperature. FIG. 10 shows the dynamic response values of S2 hybrid at situations un-bent (flat or 0) and bent to 5 ppm NO.sub.2. A response value degradation was observed at 45 (to 1.1%) and at 90 (to 1.7%) deformation, which might be attributed to the low bending energy and slight change between strained carbon atoms and NO.sub.2molecules at deformed situation. This negligible drop in response value magnitude confirms the high mechanical robustness of the fabricated sensor. In addition, no significant change in cycle-to-cycle response value (drift to 0.3%) further confirms the reliability of the sensor.

[0067] Furthermore, the fabricated sensor showed improved response-recovery time compared to reported results. FIG. 11 reveals that S2 sensor reached its maximum response value position within 7 minutes and returned to its initial position within 15 minutes. This short response-recovery time characteristics of the fabricated sensor might be accounted for by the addition of MWCNTs and RGO with WO.sub.3NPs, in which low bending energy on the sensing surface (adsorption process) and quick elimination of NO.sub.2 molecules (desorption process) are vital. However, mechanical deformation (bending angles from 0 to 90) on the sensor during sensor characterization did not show any change in the response-recovery time behaviors.

[0068] To examine the mechanical stability of the fabricated sensor, fatigue tests were carried out at different bending angles and after several times of bending-relaxing. FIG. 12(a) represents the response value variation of S2 sensor in terms of NO.sub.2 concentration at different bending angles. The sensor exhibited a limit of detection (LOD) of 1 ppm and a detection range of 1 to 25 ppm along with good linearity behavior within the entire NO.sub.2 concentration range. FIG. 12(b) shows the response value variation of S2 hybrid within 1 to 25 ppm NO.sub.2 concentration after several times of bending-relaxing processes. No remarkable degradation in response value magnitude was observed up to 10.sup.6 times of bending and relaxing processes. This can be attributed to excellent mechanical flexibility of MWCNTs and RGO, the PI substrate which was not crushed or destroyed even after 10.sup.6 times of bending and relaxing, and excellent attachment of WO.sub.3-MWCNTs-RGO network with tiny sized WO.sub.3NPs. However, after 10.sup.7 repeated cycles, the response value magnitude was dropped to about 3%, which might be caused by lower attachment between the sensing layer and the substrate due to excessive mechanical stress on the sensor.

[0069] Furthermore, humidity effect on the sensor was investigated. Humidity is one of the influencing factors on gas sensing properties (E. Traversa, Ceramic sensors for humidity detection: the state-of-the-art and future developments, Sens. Actuators B 23 (1995) 135-156; C. Wang et al., Metal oxide gas sensors: sensitivity and influencing factors, Sensors 10 (2010)2088-2106). It was reported specifically for NO.sub.2 sensing; both water and NO.sub.2 molecules behave as an oxidizing agent, and consequently, result in an increase in sensor response value (W. Yuan et al., Graphene-based gas sensors, J. Mater. Chem. A 1 (2013) 10078-10091). FIG. 13 shows the change in sensor response value of the S2 sensor (toward 5 ppm NO.sub.2) after introducing relative humidity (RH). The baseline resistance of the sensor was shifted to higher level with increasing humidity concentration. It was observed that the sensor response value was increased approximately 4% with a slight enhancement in recovery time at 81% RH.

[0070] Finally, the selectivity of the fabricated sensor (S2) was investigated by exposing the sensor to different test gases at room temperature. FIG. 14 shows the selectivity histogram of the fabricated sensor toward 5 ppm NO.sub.2 and 1000 ppm test gases, including carbon monoxide, carbon dioxide, and acetylene. As expected, S2 sensor showed better selectivity properties toward NO.sub.2, which might be attributed to superior adsorption capability of WO.sub.3 and carbon materials toward NO.sub.2 molecules.

3. Conclusions

[0071] In summary, fabrication and characterization of a high-performance NO sensor with enhanced sensing properties and excellent mechanical flexibility have been investigated at room temperature. The WO.sub.3 NPs-loaded MWCNTs-RGO hybrid at a ratio of 3:1:2 (sample S2) showed a maximum response value of 17% (5 ppm) to NO.sub.2 (a relatively short response-recovery time (7/15 minutes), an LOD of 1 ppm, and a detection range of 1 to 25 ppm. The sensor also showed exceptional mechanical flexibility and excellent repeatability at different bending angles with a negligible drift of 1.7% (at 90) and approximately 3% degradation after 10.sup.7 times of bending-relaxing processes. Less humidity effect (up to 67% RH) on the sensing surface and the excellent selectivity demonstrate that the sensor fabricated according to the present disclosure may be a suitable candidate for the fabrication of high-performance and practical NO.sub.2 sensor and be used in various sensors (vehicles, aircraft, aeronautics, and portable electronics).

[0072] As described above, preferred example embodiments of a flexible NO.sub.2 gas sensor based on a WO.sub.3 NPs coated CNTs-GO hybrid and a method for manufacturing the same according to the present disclosure are described with reference to the accompanying drawings. However, it should be noted that the present disclosure is not limited to the described embodiments and drawings but may come with a diversity of embodiments by those skilled in the art within the inventive concept of the present disclosure.